Data visualisation with predictive learning analytics

Data visualisation with predictive learning analytics

• 1.
  Data visualisation with predictivelearning analytics Chris Ballard Innovation Consultant (Analytics)
• 2.
   Background  Predictiveanalytics  Visualisation goals and issues  Examples  Guidelines Agenda
• 3.
  R&D Partnership Objective • Predictive models for student success •Map to retention themes • Visualisation Data • VLE Activity • Library Activity • Student MIS • Open data Background
• 4.
  • Retrospective • What happened? Historical •Reactive • Why? Present • Proactive • What next? Predictive Use of data in Learning Analytics When used together enables improved insight into student learning Understand student learning based on what we know now and what might happen
• 5.
  Adaptive Learning Platforms Predicting student success and atrisk students Course recommendation Using predictive analytics in education
• 6.
  Goals  Identify earlierstudents who are at risk of failure or dropping out  Understand the factors which influence student success  Simple data visualisations to help staff to support students  Actionable insights  Interventions  Monitoring Predicting student success
• 7.
  Issues with predictivemodels  They tell us what might happen, not what will happen  They are not infallible  Cannot always generate predictions  Need careful interpretation Predicting student success Appropriate visualisation is critical to its successful interpretation Predictions need to be combined with experience and knowledge of the student
• 8.
  Data visualisation examples
• 9.
  Analytics that adaptsto the user
• 10.
  Monitoring courses andmodules
• 11.
  Identifying students atrisk for a course
• 12.
  Identifying students atrisk for a module
• 13.
   Using “trafficlights” to highlight risk:  Colours can be emotive  Accessibility issues  Displaying probabilities  More vs Less granular information  Does this aid interpretation? Design considerations
• 14.
  Understand the factorswhich influence success
• 15.
  Visualisations which areeasy to interpret
• 16.
  Overlaying predictive andhistorical analytics
• 17.
  1. Visualisations shouldbe simple to interpret 2. Adapt content to the user 3. Indicate how prediction is built up 4. Bridge the gap between predictive and historic data 5. Enable users to respond and take action 6. Allow users to monitor the effectiveness of their actions Design Guidelines
• 18.
   Cross browser Responsive user interface  Support for different devices (mobile, tablet, PC)  Touch friendly Technology Guidelines
• 19.
  Thank you @chrisaballard [email protected] www.triballabs.net www.tribalgroup.com